ChatGPT Wants to Read Your Medical Records

Today AI reached into two places it had not fully touched before: your medical records and your team's daily task list. One story is about a tool that now wants to read your lab results. The other is about how little of your real work AI is actually doing, even after two years of hype. Here is what happened, and what it means for your business.

92% of VCs use AI to find companies. 58% of buyers start there, too. If you're not showing up in AI answers, you're invisible before the conversation even starts. Join HubSpot for Startups, Anthropic, and Marketing Against the Grain on July 30 (11 am ET) for a live AEO teardown. Real startup. Real recs. Register and unlock the free Startup Visibility Bundle.

Register Now

What happened: OpenAI started rolling out Health in ChatGPT to all U.S. users 18 and older on the web and iPhone, six months after a quiet pilot fell flat. You can connect Apple Health and hospital records from systems like Epic and Oracle Health, then ask questions grounded in your own medications, lab results, and visits, per OpenAI. It can compare lab results over time or summarize what changed since your last appointment. OpenAI says connected health data will not be used to train its models or target ads.

Why it matters to you: Your team is already pasting personal information into ChatGPT. A feature that pulls in real medical records raises the stakes on what people share and where it ends up. Convenience is real here, but so is the risk of trusting a confident answer about something as consequential as your health.

What to do about it: Remind your staff in one sentence that customer or employee health and personal data should never go into a public AI tool.

What happened: Google released its first ATLAS report, built from 14.6 million real interactions across Gemini and its AI search tools. It found AI now reaches 68% of occupations, covering about 90% of U.S. employment, but the average worker uses it for only about 21% of their tasks, according to Axios. Google's own economists called adoption "broad but shallow." Only 3% of jobs use AI for more than three quarters of their tasks.

Why it matters to you: The fear that AI is quietly automating away whole jobs is not showing up in the data yet. Most people use it to brainstorm, look things up, and draft, not to run the work end to end. If you have been feeling behind, that is good news. You do not need a moonshot, you need a few tasks handled well.

What to do about it: Pick one repeatable task this week, a first-draft email, a summary, or a quote, and let AI take the first pass while you keep the final say.

What happened: Patreon laid off 20% of its workforce, about 93 people, with CEO Jack Conte telling staff that "AI has fundamentally transformed the tech industry" and the company had to adjust its cost structure, TechCrunch reported. Conte was careful to say AI is not a replacement for human creativity, only that it changes how the company operates. It is Patreon's largest cut since 2022.

Why it matters to you: When a company blames AI for layoffs, it usually means two things at once: real efficiency gains, and a convenient story for cost cutting. For a small business owner, the lesson is not to slash your team. It is to notice which tasks are getting cheaper to do, and to make sure you are the one capturing that savings.

What to do about it: Before you hire for a repetitive role, ask whether a tool plus your existing team can cover it first.

What happened: Black Forest Labs, the German lab behind the FLUX image models, released FLUX 3, a model that generates 20-second videos with synced audio and, more surprisingly, controls robots, per VentureBeat. A spinoff called FLUX-mimic is already being tested on Audi's production line, learning a new task from about 30 minutes of demonstration instead of the usual 30 hours or more. The company's bet is that the same "visual intelligence" that makes video also lets machines act in the real world.

Why it matters to you: This is a preview of where the next wave lands, not just on screens, but in physical work. It is early and limited, so there is no reason to panic. But the line between AI that makes content and AI that does labor is starting to blur, and that is worth watching if your business runs on hands-on tasks.

What happened: Anthropic upgraded Claude's voice mode so paid users can hold spoken conversations on its smarter Opus and Sonnet models, not just the basic one, TechCrunch reported. More useful for business owners: you can now trigger actions by voice in Gmail, Google Calendar, Slack, Canva, and Notion. Free users get the basic model and one connected app.

Why it matters to you: Voice is quietly becoming a real way to get work done, not a gimmick. If you spend your day in email, calendar, and a couple of apps, talking through a task while you drive or walk starts to make sense.

What to do about it: Try it on something low-stakes first, like drafting a calendar invite, before you trust it with anything that matters.

Two things are true after today. AI is reaching further into your health, your apps, and even your factory floor. And it is still doing a smaller share of the actual work than the headlines want you to believe. That gap is where calm business owners win. You do not have to bet the company or fire your team. You have to pick a few tasks, hand them off carefully, and keep your hand on the wheel. That is how you ride this thing instead of getting thrown off it.

If today's issue made the noise a little quieter, forward it to one business owner who needs it. New here? Subscribe and ride along with us every morning.

Talk tomorrow,
Mark Shilensky

Follow me personally on Social Media:
Facebook: https://www.facebook.com/MarkShilenskyPage
Instagram: https://www.instagram.com/markashilensky/
Threads: https://www.threads.com/@markashilensky
LinkedIn: https://www.linkedin.com/in/shilensky/
X: https://x.com/markshilensky
TikTok: https://www.tiktok.com/@mark.shilensky
BlueSky: https://bsky.app/profile/markshilensky.bsky.social